This release replaces antigravity-preview-05-2026. It introduces changes to built-in tools, including PascalCase parameters and line-range replacements for file edits.

ai.google.dev ↗

This high-reasoning audio-to-audio model supports background reasoning during live audio interactions. It is recommended when higher background reasoning is required.

ai.google.dev ↗

It is the default option for most low-latency voice agent experiences and real-time dialogue without reasoning delays. Features interleaved reasoning, default asynchronous function calling, and full session client content updates.

ai.google.dev ↗

GPT-6 Astra is designed for reasoning, coding, computer use, research, and document creation. It does not support custom temperature or top_p values, and tool calling requires the Responses API.

platform.openai.com ↗

Optimized for 30-second clips, loops, and rapid experimentation. It supports text and image inputs and generates high-fidelity 44.1 kHz stereo audio.

ai.google.dev ↗

Optimized for full-length song generation with improved musical coherence, natural vocals, and fine-grained duration and structural control. It supports text and image inputs and generates high-fidelity 44.1 kHz stereo audio.

ai.google.dev ↗

It is engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows.

ai.google.dev ↗

It is the successor to Claude Fable 5 for long-running agentic coding, knowledge work, and research. It supports always-on adaptive thinking and is available on the Claude API, Amazon Bedrock, AWS, Google Cloud, and Microsoft Foundry.

docs.claude.com ↗

Released gemini-omni-1.1-flash, the GA version of the fast, conversational video generation and editing model. This release includes new capabilities like video extension, interpolation, and resolution control up to 4K.

ai.google.dev ↗

It offers low-latency, bidirectional streaming speech-to-text over WebSockets using the Live API. It supports interim and finalized transcription events, Smart transcription mode, and multiple Voice Activity Detection (VAD) strategies.

ai.google.dev ↗

It is a high-accuracy, low-latency non-streaming speech-to-text model with utterance-based language detection across 85+ languages. It supports speaker diarization, word-level timestamps, and custom vocabulary biasing up to 1,000 terms.

ai.google.dev ↗

Released as our most intelligent workhorse model yet for coding and agents, with substantial improvements across software engineering, web development, and agentic workflows. It is available at an introductory price through December 31, 2026.

ai.google.dev ↗

Available in beta on Claude Fable 5, Claude Mythos 5, Claude Opus 4.8, and Claude Opus 5, allowing adding or removing tools between turns while preserving the prompt cache via the mid-conversation-tool-changes-2026-07-01 beta header.

docs.claude.com ↗

The model supports the full effort ladder (low, medium, high, xhigh, max), with max for capability-critical work.

docs.claude.com ↗

Using thinking disabled with effort xhigh or max returns a 400 error, a breaking change from Claude Opus 4.8.

docs.claude.com ↗

Dreams (research preview) now supports Claude Opus 5. Details are in the Supported models documentation.

docs.claude.com ↗

Supports a 1M token context window (both default and maximum), 128k max output tokens, and thinking on by default, priced at $5/$25 USD per MTok, the same as Claude Opus 4.8. Available on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry.

docs.claude.com ↗

Optimized for real-time text streaming using the Live API, enabling low-latency robot agents with bidirectional audio and video input. Accepts text, image, video, and audio inputs and supports function calling with blocking behavior.

ai.google.dev ↗

Embodied reasoning endpoint offering advanced spatial reasoning, agentic code execution, multi-step tool orchestration, video moment finding, progress classification, and multi-robot coordination. Accepts text, image, video, and audio inputs and supports function calling with blocking behavior for physical robot actions.

ai.google.dev ↗

Provides low-latency streaming transcription. Supports free-form transcription context, keyword hints, and multiple expected input languages.

platform.openai.com ↗

Supports free-form transcription context, keyword hints, and multiple expected input languages. It pairs with the v1/audio/transcriptions API and v1/realtime for final transcripts of committed Realtime turns.

platform.openai.com ↗

Released as a stable, production-ready version offering a low-latency, highly cost-effective subagent option designed for high-volume automation.

ai.google.dev ↗

Released as a stable, production-ready version with improved token efficiency and code/agentic planning capabilities at a lower price point than 3.5 Flash, resolving developer feedback around output verbosity.

ai.google.dev ↗